Automated Classification and Valuation of Ceramic Artifacts Using Deep and Machine Learning Models

This study addresses the challenge of automating ceramic artifact classification and valuation, which has traditionally relied on expert assessments, often introducing subjectivity and inconsistency. By leveraging advanced machine learning and deep learning techniques, the research utilizes the YOLOv11 model to classify key ceramic attributes such as decorative patterns, shapes, and craftsmanship style. A Random Forest classifier is then used to predict the market value of ceramics based on features extracted by YOLOv11, alongside auction data. The model achieved significant improvements, with an mAP@50 of 57%, recall of 91%, and an overall accuracy of 94.36% for price classification. Feature importance analysis revealed that manufacturing techniques played a crucial role in determining ceramic prices. This study demonstrates the effectiveness of combining deep learning for feature extraction and machine learning for price prediction, providing a reliable and automated tool for ceramic classification and valuation, with significant implications for both academic research and practical applications in artifact appraisal and museum digitization.